Instinct

# Fix list: Instinct

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/instinct, the October 2026 research run, assessed 12 October 2026. Grade F, 11.4 out of 100.

This is everything the published grade says the listing lacks, the biggest possible gain to the total first. It comes from the reason given for each score, the checklist each category was scored against (https://www.anchorterminal.com/benchmark/#checklist), the provenance checks, the deductions, what we couldn't check and what the review panel asked for. A fix counts at the next check, once it's public.

For a coding agent working on Instinct: work through the items below in the product, its docs and its public pages. Each category gives the reason for its score, with the points each checklist item earned, and the checklist itself, so the gap is the items that earned less than their points. Change the product, not the wording, and keep a note of what you changed and where it's published.

## 1. Reliability, 5 out of 100, up to 19 more on the total

Why it scored 5: Read as a hosted service. No status page is linked from the home page, terms or privacy policy (0), so there is no incident record to read (0). No rate or usage limits are published (0), and nothing on retries or overload (0). No SLA was found, and the terms supply the service as is and as available and say access may be interrupted (0). The service is live with a sign-in and no beta label on the pages read, but the SMS terms describe a waiting list when it is at capacity and nothing declares general availability (5 of 10).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-reliability):

Hosted APIs, MCP servers, models and platforms.

- 20, a public status page with component history (Statuspage, Instatus, BetterStack or the vendor's own).
- 0 to 30, the incident record for the last 90 days on that page. 30 for a clean record or trivial incidents only, 20 for minor incidents only, 10 for one major outage (an hour or more of a core API down, or errors across the board), 0 for several. 5 when there's no history we could read, and the note says so.
- 15, rate limits documented with numbers.
- 15, documented 429 or overload handling (Retry-After, backoff guidance), and idempotency keys or safe-retry guidance where writes are involved.
- 10, an SLA published for any paid tier.
- 10, the surface agents use is generally available, not beta or preview.

Local packages, SDKs, frameworks and stdio MCP servers.

- 20, installs from an official package with supported runtimes stated.
- 25, a public CI and test suite, passing on the default branch.
- 0 to 25, open crash or regression issues relative to activity (25 for few and handled, 0 for many, old and unanswered).
- 15, semver discipline and breaking changes called out in a changelog.
- 15, version 1.0 or later, or declared stable.

Protocols are read from their reference implementations, the public facilitators or servers, spec stability and test vectors.

## 2. Schema & documentation, 3 out of 100, up to 15.8 more on the total

Why it scored 3: Instinct publishes no API, MCP server or other interface a program can call, so there is no machine-readable contract (0). `/llms.txt` answers with the web app's page, not a text file (0). The public description is five sentences on the home page plus the legal documents, which say what the assistant may do and with which accounts (3 of 20). No typed inputs (0), no examples or documented errors (0), and no versioning or changelog (0).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-schema):

APIs and MCP servers.

- 25, a machine-readable contract (a public OpenAPI file or similar; for MCP, typed JSON Schema inputs on every tool).
- 10, llms.txt or Markdown docs served for agents.
- 0 to 20, descriptions that say what a tool is for, when to use it and when not to, read from the tool definitions in the source or the API reference.
- 0 to 15, typed inputs with enums, constraints and required fields, and no free-form JSON blobs.
- 0 to 15, examples and documented error responses.
- 15, versioning and a public changelog.

Models are read from the API reference, the OpenAPI file, llms.txt, the structured-output and tool-use docs and the model cards. Frameworks from docs a model can follow, typed interfaces, examples and the API reference.

## 3. Agent ergonomics, 5 out of 100, up to 15.4 more on the total

Why it scored 5: There is no programmatic interface to size, page, retry or annotate, so the first four lines score 0. The owner needs no client or syntax, since requests go by text message or phone call in plain language, and there is no SDK (5 of 15).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-ergonomics):

- 0 to 25, context cost. For MCP, the number and size of the tool definitions (25 for ten or fewer compact tools, 15 for 11 to 30, 5 for more than 30, plus up to 10 back for toolsets, dynamic loading or read-only subsets). For APIs, whether responses can be sized (field selection, limits, summaries).
- 20, pagination, filtering and output-size controls.
- 20, actionable, documented error responses, codes and messages an agent can recover from.
- 20, idempotency or safe retries, and for MCP the `readOnlyHint` and `destructiveHint` annotations.
- 15, sensible defaults, few required parameters, and official SDKs in at least two languages.

Models are read for tool use, structured output, prompt caching, context length, batch and SDKs. Frameworks for how much code and how many defaults a tool-calling agent with MCP needs.

## 4. Security & auth, 25 out of 100, up to 13.1 more on the total

Why it scored 25: The owner signs in with a one-time passcode sent by text, and no second factor is described. Google Workspace is connected through Google's APIs and can be revoked, but the access listed spans Calendar, Gmail, Drive, Docs, Sheets, Slides and Tasks with no scopes published, and the privacy policy says the owner may hand over usernames, passwords and card details, with storage not described (10 of 30). The terms say confirmation requirements may apply to certain actions, with no list, no setting described and no warranty, while the assistant may make purchases and accept third-party terms for the owner (4 of 20). The privacy policy names misleading instructions from third parties as a risk and asks the owner to review actions, with no mitigation described (3 of 15). The terms refer to records of actions in the service and say they may not always be accurate (4 of 15). security.txt is valid until 18 February 2027 with a contact address and no policy. No bounty, certification or published advisory was found. The policy commits to Google's Limited Use requirements (4 of 20).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-security):

- 0 to 30, the credential model. 30 for OAuth 2.1 with scopes, or scoped and revocable keys with rotation. 20 for plain revocable API keys. 10 for one all-powerful key. 10 off when a secret can travel in a URL query string as a documented option.
- 0 to 20, read-only or least-privilege modes, and confirmation or approval for destructive actions.
- 0 to 15, prompt-injection posture where the tool returns untrusted content (documented mitigations or guidance). A tool that returns no untrusted content gets 10.
- 0 to 15, audit logs or per-call visibility for the operator.
- 0 to 20, a security programme. security.txt or a disclosure policy, a bug bounty, SOC 2 or ISO 27001, advisories handled in public.

Models are read for retention, whether API data trains models (and whether that's off by default), zero-retention options and certifications. Frameworks for telemetry defaults, approval hooks, guardrails and sandboxing.

## 5. Payments & pricing, 0 out of 100, up to 12.5 more on the total

Why it scored 0: No x402, MPP or L402 (0). No price is published. The terms say only that fees may apply to certain functions and are non-refundable (0). A free tier or trial without a card was not established from the public pages (0). A person signs in with a phone number in a browser and may be placed on a waiting list (0).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-payments):

The published rubric, also on the [x402 page](https://www.anchorterminal.com/x402/).

- 40, a machine payment protocol (x402, MPP or L402) on the tool's own endpoints. 10 to 30 when it covers only some endpoints or only goes through a third party, and the note says which.
- 20, per-call or per-unit pricing published without a login. 10 for public plan-only pricing, 0 for "contact sales" or prices behind a login.
- 20, a free tier or trial that doesn't need a card.
- 20, autonomous onboarding, meaning an agent can get access without a person signing up in a browser (keyless use, x402, a programmatic key API).

Payment platforms and agent wallets rarely charge for their own API over a machine protocol, so the first line has steps for them, and the highest one that applies counts. 40 when x402, MPP or L402 runs on all their own endpoints, 30 when it runs on part of their own API, 25 when their merchants can accept one, 20 for running a facilitator, 15 for paying as a buyer, and 0 when the only protocol is their own. Merchant acceptance sits above a facilitator because the platform's own customers can charge agents through it, while a facilitator settles for sellers who wire up the protocol themselves. The counter-argument (a facilitator does more for the protocol as a whole) has a point. Each note says which step applied.

Open-source software you run yourself is scored on its hosted or paid option if it has one. A free, self-hosted package with nothing to buy gets 20, 20 and 20 for the last three lines, and 0 to 40 for the first only if it ships a payment protocol.

## 6. Maintenance & community, 3 out of 100, up to 8.5 more on the total

Why it scored 3: No changelog, release notes or dated product announcement was found, so recency scores 0 and there are no dated entries in the last 90 days (0). The newest dated documents are the terms and privacy policy of 26 August 2026, which are legal revisions, not releases. Support and privacy email addresses are published, with no community channel or public tracker found (3 of 15). No SDK, package or registry entry (0), and no package to assess (0).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-maintenance):

- 0 to 30, time since the last release, or the last published model or API change for a closed service. 30 within 30 days, 20 within 90, 10 within 180, 0 older.
- 20, at least three releases or dated changelog entries in the last 90 days.
- 0 to 25, responsiveness. Issues and pull requests answered on GitHub (the open issues and how recent the replies are). For closed services, a public changelog and a support or community channel that answers, 0 to 15.
- 15, presence in the official MCP registry under a verified namespace (MCP servers), or current official SDKs (APIs and models).
- 10, package health, current dependencies and CI.

Models are read for deprecation notice periods and model churn rather than release counts.

## 7. Transparency & trust, 51 out of 100, up to 4.3 more on the total

Made of editorial 30, provenance 72.

Why it scored 51: Closed service with published terms of 26 August 2026 that name the legal entity and cover the assistant, the apps and the website (15 of 30). The privacy policy says what is collected from Google Workspace, that customer content trains AI models unless the owner opts out, that Workspace data and Vault material are excluded, and how indexed data is deleted. It gives no retention period, no DPA was found, and the terms say the company may but need not delete materials when an account is deleted, which sits uneasily with the policy's deletion route (12 of 30). No deprecation policy. The terms let the vendor change or remove any part of the service at any time (0 of 20). Recipients are given by category (AI model providers, a payment processor, hosting) with no names, and the terms say the service is operated in the United States (3 of 20).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-transparency):

- 0 to 30, source availability and licence clarity. 30 for open source under an OSI licence, 15 for closed with clear terms, 0 for unclear terms.
- 0 to 30, data handling and retention statements that agree with each other (privacy policy, DPA, retention periods, subprocessors).
- 0 to 20, a deprecation policy or notices with dates.
- 0 to 20, telemetry disclosed with an opt-out (local software), or subprocessors and data locations disclosed (hosted).

The other half of Transparency and trust is the provenance score, computed from checked facts (below). The category score is the mean of the two.

Provenance checks not met in full (half of this category, computed from checked facts):

- Terms of service: read, states 6 of the 7 things a reader expects, and has 2 clauses that cost points (5.1 of 10)
- Privacy policy: read, states 4 of the 8 things a reader expects (7 of 10)
- Status page: not found (0 of 10)
- Changelog: not found (0 of 10)

## What we couldn't check

What we couldn't read counted as absent. Publishing it on a page a plain HTTP fetch can read (not only in a browser) lets the next check count it.

- Kind is `platform`. The briefs define no kind for a personal assistant, and they give `platform` for a web, desktop or phone app, which is how Poke is recorded. Instinct has no API, so `http-api` does not apply. Graded as a hosted service.
- Instinct has no interface an outside program can call. The files were written because the category covers hosted assistants reached by text, but Schema, Ergonomics, Payments and Maintenance are near 0 for lack of any public page to score, and the listing may be better held.
- The address in the lead is right and the product is a personal assistant. The lead's description matches the home page read on 2026-10-12. It said the assistant is started by text from the website. The home page button named Text Instinct links to the sign-in at app.instinct.com.
- `pricing` is recorded as `paid` because the terms describe fees and the policy describes card payments. No price is published, and whether a free tier exists was not established.
- `assistant.proactive` rests on the privacy policy calling the assistant autonomous and on its mention of suggestions. `assistant.memory` rests on the policy's wording about indexing connected data and drawing on earlier interactions. Neither is described in product documentation.
- The terms (section 6) forbid automation software, robots, scrapers and benchmarking and limit use to personal use. Recorded as a fact with no deduction. It matters before any probe is run.
- unchecked: the signed-in app at app.instinct.com, including the sign-in page, settings and workspace pages, which are script-drawn. Approval controls, the action log, the Vault, the list of connectable services and any price shown after sign-in were not read.
- unchecked: which model suppliers, hosting company and payment processor see the content. None is named on the pages read. A response header from app.instinct.com permits scripts from Cloudflare, Duffel and Evervault, which is a hint and not a statement of who processes data.
- unchecked: Google OAuth scopes, where tokens, passwords and card details are stored, and whether they are encrypted.
- unchecked: retention periods and what account deletion removes. The policy and the terms differ on whether materials are deleted.
- unchecked: the Mac OS and mobile apps. No app store link appears on the pages read, so their existence rests on the terms and privacy policy.
- unchecked: a GitHub organisation or package for the vendor. A search for Spear Street Technology returned none, and the many accounts with instinct in the name could not be tied to it. The MCP registry entries matching instinct belong to other publishers.
- unchecked: security advisories and incidents. None is published on the vendor's pages and no security page exists to carry one.
- instinct.com/robots.txt answered 200 and allows every agent. app.instinct.com/robots.txt answered with the app's page, read as no rules, and nothing else was requested from that host. rdap.verisign.com answered 400 and api.github.com and registry.modelcontextprotocol.io 404 to the robots.txt request, read as no rules.

## Weaknesses

- No price is published. The terms say fees may apply to certain functions, are charged in US dollars and are non-refundable
- No documentation, API, MCP server, status page or changelog was found on instinct.com
- Customer content trains the vendor's AI models by default, and content flagged for safety review may be used after an opt-out
- The terms appoint the assistant as the owner's agent to make purchases and accept third-party terms, and leave responsibility for every action with the owner
- The owner may hand over usernames, passwords and card details. Where they are stored is not stated
- The terms forbid bots, scrapers and benchmarking, and limit the service to personal use

## What costs an agent a turn today

The notes we give agents before they call it. Each one is a workaround an agent shouldn't need.

- Do not plan on calling Instinct from a program. No API, MCP server or key was found, and the terms forbid automation software and benchmarking
- A person signs in at app.instinct.com with a phone number and a one-time code by text, and may wait for a place when the service is at capacity
- Tell the owner to opt out of model training at app.instinct.com/settings before connecting mail or files, since training is on by default
- Disconnecting an account does not delete what was indexed from it. Delete that data at app.instinct.com/workspace
- Treat purchases and bookings as final. The terms say actions may not be reversible and that confirmation controls are not guaranteed

## When it's done

Send what changed and where it's published as a dispute (https://www.anchorterminal.com/builders/#disputes, or `POST https://www.anchorterminal.com/api/v1/contact` with `"kind": "dispute"`). Disputes are answered in public, and the listing is checked again by the same checklist. Paying for an audit or a listing claim changes nothing here.